jmedev environment

[riccardobl]

The goal of this PR is to improve the experience of setting up a local or remote development environment to start contributing to JME. This is achieved by fixing #2626, improving some configurations, and providing a set of devcontainers.

These containers can be run from Visual Studio Code or from any environment that supports Podman or Docker.

For Linux + Wayland machines

This is the best option, as these containers can run at native performance. Thanks to Wayland, the JME window is presented on the host, making it virtually indistinguishable from running it directly on the host system.

• JME env: Runs with GPU passthrough where possible, or falls back to software rendering.
• JME env (NVIDIA): Similar to JME env, but specifically configured for NVIDIA GPUs.

For everything else

• JME env (SERVER): This container has no GPU passthrough whatsoever and runs fully software-rendered on a virtual VNC server exposed via noVNC.

It can run in environments where the other containers cannot, including servers, virtual machines, docker on windows and mac, GitHub Codespaces, which can be very useful for a quick PR review or to test a branch. However, performance will be significantly lower.